I am proposing to research the reasons behind the deliberate selection of postmodern architectural materials when crafting cultural landmarks because I want to find out how they are used as a diplomatic tool of soft power. This will help my reader understand how architecture is harnessed to evoke a national identity on the world stage. Thus, my research question is: what explains variation in the use of postmodern architectural materials when crafting globally visible buildings?
It would be most useful to create my own dataset for this research question. The dependent variable is architectural material, which would be operationalized as the frequency with which the material appears in a large number of buildings. The buildings are the cases. One option is to measure how much of the building is composed of a particular postmodern material (buildings that are 95% steel, an industrial material, and 5% colored glass, a postmodern material, for example, probably wouldn’t tell us much about postmodern influence, so it may be prudent to make the distinction), but because data of this sort is scarce, it makes more sense to simply measure how often these materials appear across many architectural structures. The dependent variable could take the form of stone, colored glass, ceramics, or cantilevered beams, and the level of measurement would be the percent of buildings composed partially or fully of this material–thus, the level of measurement would be nominal. I could also–I believe–group certain buildings into categories of percentages (0-25% postmodern material, 26-50%, etc.), to make it an interval level of measurement.
What I am hoping to do is consider all of these materials- glass, ceramics, etcetera- as mere variations on one dependent variable (postmodern material), rather than multiple dependent variables, although I am unsure if this is possible.
One existing dataset I would use to compile this information would be the Historic American Buildings Landscape dataset, which I accessed through ICPSR, which gathered information from blueprints, photographs, and written descriptions of American architectural spaces to evaluate the color choice and material choice used in each. This would not be sufficient on its own, of course—I would, ideally, like to find similar data sets for multiple regions of the world.
The dataset uses 40,000 buildings and structures, and, among other information, lists the materials of which they are composed, measured qualitatively in large groupings like “glass,” “concrete,” “steel,” and “wood.” I would operationalize these by measuring, once again, the frequency with which the materials—specifically the materials identified by scholars as postmodern—appear, and then transform those into percentages. For example, I would determine a way to measure all of the times that “glass” appears as a material, and begin to ask questions such as, “Glass appears 40% of the time in American postmodern spaces. What is alike or different about those spaces (independent variables) that may have caused this choice to be made?”
The dataset is limited both temporally and geographically: it doesn’t cover anything after the year 2000, and although it covers many buildings from the 1500s and onwards, I cannot use most of these, as they would not be classified as ‘postmodern.’ Thus, I am limited to using the data from 1960-2000, which, although still a wealth of data, is not nearly as rich as the 40,000 potential structures covered by the entire data set. The dataset also focuses specifically on American postmodern spaces, and while I may recognize later on that I need to confine my statistical research design to a particular region of the world, I have not settled on one yet, and therefore, I would need more datasets to compile this information. If postmodernism is especially rich in Western Europe, for example, I would rather pivot to that region and search for data collected on postmodern materials there (although I struggled to find this data for this particular post, which tells me it may not be out there). Finally, as aforementioned, the dataset is limited in that it doesn’t provide any information on how much of the building is composed of that particular material, which I would find more helpful and thorough than just knowing that the material is present.
 “Historic American Buildings Survey/Historic American Engineering Record.” Washington, D.C., U.S. Dept. of the Interior, National Park Service, Cultural Resources, 1994. 5, 45.
 Ibid, 5.